SOA Enables Outsourced Imaging

Service-oriented architecture (SOA) offers an opportunity for radiologists to access multiple services off the internet and may be a more economical option for hospitals. While some PACS vendors see it as a threat, others view SOA as an opportunity to offer more services.
A leading advocate for healthcare’s adoption of SOA, Stephan G. Erberich, assistant professor of Research Radiology and Biomedical Engineering, The Saban Research Institute, Children’s Hospital Los Angeles, explains how SOA is breaking up the traditional concepts of what it means to be a PACS.
The SOA provides an opportunity for hospitals to consider the recipe for services needed to be a functional radiology department. There is already decoupling of what is an internal and external service - teleradiology is one example. Many hospitals already use outside radiologists, which is a SOA – looking for outside services and decoupling services into individual ones.
For radiology, SOA changes the way in-house resources have to be looked at, and it creates a huge dependency, as a trade off, on the network environment. Most hospitals usually have a T1 line, but they don’t utilize that as part of their operational component. The concept of SOA can change that paradigm.
Erberich illustrates how SOA-based image processing could work in the clinical setting. Hospitals today have to buy very expensive image processing capacities and the utilization of these resources is usually very limited. In order to utilize that to more efficient levels, it would be better to share this environment for doing the image processing. The problem is that you need to provide this service on the Internet so that it is reachable by several hospitals.
A cardiology reconstruction service for heart rendering could be provided as a service on the Internet. Hospitals would send the data to that service provider and get these images processed and returned. That is more attractive for hospitals because then they only have effective costs. It depends on utilization. It’s very expensive to have high-upfront costs in order to bring this very specialized equipment in, and if there are only a few cases it is hard to justify the usefulness of that functionality. With an SOA that outsources that specific function, you have the benefit of providing the best care to the patient at any facility and you are utilizing the equipment and the cost associated with that equipment, in this case image processing, to the maximum degree. So the advantage of SOA is it defines the standard communication protocol used. In this case, SOA messages, which are services interface definitions, actually provide a standard interface layer even for proprietary hardware.
If you have a fMRI data set and you want to get it processed, you can send the DICOM images out into a shared environment. Grid technology is one of the infrastructures that supports the SOA concept. You send the DICOM into the “grid cloud.” There you will find your service provided by a commercial entity. If you are a subscriber to that service, using those standard protocols in the grid layer, you send your images to get processed and returned to you. This allows you to completely outsource that specific function, and it makes this function available virtually at any institution that creates these types of images.